Seismo-hydro-mechanical modelling of the seismic cycle: Methodology and implications for subduction zone seismicity

Abstract

Slip accommodation in subduction zones ranges from aseismic slip phenomena to regular megathrust earthquakes, and strongly depends on pore fluid pressure. We develop a new fully coupled poro-visco-elasto-plastic seismo-hydro-mechanical numerical model, allowing for coupled modelling of tectonic and seismic processes in the presence of fluids. A combination of fully staggered finite differences and marker in cell techniques is used to solve mass and momentum conservation equations for solid matrix and fluid coupled to a poro-visco-elasto-plastic rheological constitutive relationship. Brittle/plastic deformation is resolved through global Picard-iterations and adaptive time stepping is introduced to resolve time scales from milliseconds to thousands of years involved in the hydro-mechanical seismic cycle. We demonstrate how and why the presence of pervasive fluid flow in the deforming poro-visco-elasto-plastic subduction interface causes localisation of deformation and nucleation of seismic events with slip rate up to m/s. The nucleation of fault slip is controlled by rapid fluid pressure increase due to visco-plastic compaction of a spontaneously forming fault balanced by the simultaneous elastic decompaction of deforming pores inside the fault. Subsequent post- and inter-seismic slow fluid pressure release by elastic compaction of the stressed pores allows recovery of subduction interface strength. The events nucleate downdip in the brittle-ductile transition zone and propagate updip in form of highly localized, spontaneous ruptures. The model reproduces the broad spectrum of transient phenomena ranging from slow slip to seismic ruptures on a subduction interface with homogenous elastic and frictional properties that do not depend on slip rate. The degree of locking of the megathrust interface, the coseismic stress drop and the dominant slip regime during subduction are critically dependent on effective large-scale and long-term rock permeability. A decrease of permeability leads to a decrease of degree of locking, leading to smaller stress drop and enhancing the occurrence of stable aseismic slip.